Tom McCoy

@rtommccoy.bsky.social

2018 Followers 347 Following

Assistant professor at Yale Linguistics. Studying computational linguistics, cognitive science, and AI. He/him.

Schedule for some presentations at CoNLL: 2:00 - 3:30 poster session in Harbor C: "What exactly do children receive in language acquisition? A case study on CHILDES with automated detection of filler-gap dependencies" 4:00 - 5:30 talk session in Harbor C: "Collocational bootstrapping: A hypothesis about the learning of subject-verb agreement in humans and neural networks"
Screenshot from the New York Times crossword page that says "The Crossword. Monday, June 8, 2026. By Tom McCoy. Edited by Will Shortz."
Screenshot of an article title reading "The 5 Jane Austen books everyone should read for what would be her 250th birthday"
A Zipfian distribution fitted to child-directed language; the empirical distribution matches the theoretical one closely
Plot showing neural net subject-verb agreement accuracy as a function of the variability in the training data. Accuracy is optimized (and is near 1.0) when the level of variability is medium.
Zipfian distributions with various values of the alpha parameter that modulates how skewed the distribution is
Paper overview. Title: "Collocational bootstrapping: A hypothesis about the learning of subject-verb agreement in humans and neural networks" Authors: Claire Hobbs and Tom McCoy Method: We trained many neural nets, varying how predictable a subject is given its verb. We tested them on subject-verb agreement Findings: With the right level of predictability, neural networks robustly generalize. The predictability of child-directed language is near the neural net optimum. Conclusion: Statistical regularities in word co-occurrence can support the learning of abstract syntactic rules The text is accompanied by a graph showing neural-network accuracy as a function of the level of variability; the accuracy peaks at an in-between level of variability
Screenshot of the title and abstract of a paper. Title: "Not-So-Strange Love: Language Models and Generative Linguistic Theories are More Compatible than They Appear" Further information: "Open Peer Commentary on “How Linguistics Learned to Stop Worrying and Love the Language Models” by Richard Futrell and Kyle Mahowald" Author: R. Thomas McCoy Abstract: Futrell and Mahowald (2025) frame the success of neural language models (LMs) as supporting gradient, usage-based linguistic theories. I argue that LMs can also instantiate theories based on formal structures - the types of theories seen in the generative tradition. This argument expands the space of theories that can be tested with LMs, potentially enabling reconciliations between usage-based and generative accounts.
An image of two snack foods. At the top is Reese's Oreo, which is an Oreo-flavored brand of Reese's. At the bottom is Oreo Reese's, which is a Reese's-flavored brand of Oreo.
Image of a Laffy Taffy wrapper. The joke asks "What do you deserve and is also a type of bagel?" And the answer is "Everything"
Top: A syntax tree for the sentence "the doctor by the lawyer saw the artist". Bottom: A continuous vector.
Illustration of the blog post's main argument, summarized as: "Theory of Mind as a Central Skill for Researchers: Research involves many skills.If each skill is viewed separately, each one takes a long time to learn. These skills can instead be connected via theory of mind – the ability to reason about the mental states of others. This allows you to transfer your abilities across areas, making it easier to gain new skills."
On the left is the cover of David Copperfield by Charles Dickens. On the right is the cover of Demon Copperhead by Barbara Kingsolver.
The top shows the title and authors of the paper: "Whither symbols in the era of advanced neural networks?" by Tom Griffiths, Brenden Lake, Tom McCoy, Ellie Pavlick, and Taylor Webb. At the bottom is text saying "Modern neural networks display capacities traditionally believed to require symbolic systems. This motivates a re-assessment of the role of symbols in cognitive theories." In the middle is a graphic illustrating this text by showing three capacities: compositionality, productivity, and inductive biases. For each one, there is an illustration of a neural network displaying it. For compositionality, the illustration is DALL-E 3 creating an image of a teddy bear skateboarding in Times Square. For productivity, the illustration is novel words produced by GPT-2: "IKEA-ness", "nonneotropical", "Brazilianisms", "quackdom", "Smurfverse". For inductive biases, the illustration is a graph showing that a meta-learned neural network can learn formal languages from a small number of examples.
Picture of a paragraph from Emma by Jane Austen. It reads: "Such an adventure as this, a fine young man and a lovely young woman thrown together in such a way, could hardly fail of suggesting certain ideas to the coldest heart and the steadiest brain. So Emma thought, at least. Could a linguist, could a grammarian, could even a mathematician have seen what she did, have witnessed their appearance together, and heard their history of it, without feeling that circumstances had been at work to make them peculiarly interesting to each other? How much more must an imaginist, like herself, be on fire with speculation and foresight? especially with such a groundwork of anticipation as her mind had already made."
Left: A plot showing recursion results for standard and prior-trained neural networks. The x-axis shows levels of recursion ranging from 0 to 10, and the y-axis shows accuracy. As the levels of recursion increase, the accuracy drops for both models, but it drops much more rapidly for the standard model than the prior-trained model. Right: A plot showing priming results. There are 4 sub-plots, for 4 types of sentences: short plausible, long plausible, short implausible, and long implausible. In all 4 plots, the prior-trained network shows a greater degree of priming than the standard neural network.
A plot showing perplexity values. Note that for perplexity, lower is better. A standard neural network achieves perplexity ranging from about 19.70 to 19.80, with a median around 19.75. A prior-trained neural network achieves perplexity ranging from about 19.63 to 19.74, with a median around 19.67. The best model from prior literature is indicated as having a perplexity of about 19.69.
Plots showing results for formal languages. On the left is a line graph which has “number of training examples” as its x-axis and “F-score” as its y-axis. Three models have lines in this plot: a Bayesian model, a standard neural network, and a prior-trained neural network. The Bayesian model and prior-trained neural network perform similarly, while the standard neural network does much worse than both of them. On the right is a table showing the amount of training time used by each approach. The Bayesian model uses from 1 minute to 7 days of training time. The neural networks (whether standard or prior-trained) use from 10 milliseconds to 2.5 minutes.
Two examples of formal languages. The first example shown is AnBn, described as “n copies of A followed by n copies of B”, with some example strings from the formal language being AB, AABB, and AAABBB. The second example shown is XXX, described as “any string X repeated three times”, with some example strings from the formal language being AAA, BABABA, and ABBABBABB
A video using two graphs to show meta-learning using MAML. The model is a simple linear function with 2 parameters (slope and intercept): a star in the left plot shows the parameters, and a solid line in the right plot shows the function defined by these parameters. This model is exposed to many target tasks, each a linear function defined by a square in the left plot. All of these squares are drawn from the perimeter of a circle, providing some similarity across tasks that meta-learning can exploit. For each target task, a copy of the model is trained on 10 examples (shown as points in the right plot), yielding the parameters shown with an orange circle in the left plot and a function shown by a dotted line in the right plot. Based on how close this copy is to the target, the model's initialization is adjusted. After many iterations, the model's initialization reaches a point from which the model can learn any target task—somewhere in the middle of the circle in the left plot.
A schematic diagram of our procedure. We start with a Bayesian model, here visualized with Bayes’ rule and some example grammatical rules that could be sampled from a Bayesian model’s prior. Then, we sample several tasks from that Bayesian model’s prior, which can serve as training data. Finally, we have a neural network meta-learn from these sampled tasks. The whole process is visualized, going from left to right as “Bayesian model’, then an arrow labeled “sampling”, then “training data”, then an arrow labeled “meta-learning”, and finally “neural network.”
Left: A screenshot of ChatGPT describing itself as an AI language model developed by OpenAI. Right: A bar chart comparing the quantity of text seen by human children vs. GPT-3. The bar for GPT-3 is far higher than for humans, showing that neural networks get far more linguistic data than humans do.
Screenshot of a demo of Bayesian word learning (Xu & Tenenbaum 2007). After a few examples, the Bayesian learner figures out that naysayer means “horse” (rather than being more specific – “horse number 4” – or more general – “mammal”).
A schematic of our method. On the left are shown Bayesian inference (visualized using Bayes’ rule and a portrait of the Reverend Bayes) and neural networks (visualized as a weight matrix). Then, an arrow labeled “meta-learning” combines Bayesian inference and neural networks into a “prior-trained neural network”, described as a neural network that has the priors of a Bayesian model – visualized as the same portrait of Reverend Bayes but made out of numbers. Finally, an arrow labeled “learning” goes from the prior-trained neural network to two examples of what it can learn: formal languages (visualized with a finite-state automaton) and aspects of English syntax (visualized with a parse tree for the sentence “colorless green ideas sleep furiously”).
Screenshot of the New York Times crossword page, saying "The Crossword. Wednesday, May 7, 2025. By Tom McCoy. Edited by Will Shortz."
On the left is a probabilistic context free grammar (PCFG). On the right is an image of the Transformer architecture. There are arrows going back and forth between the PCFG and the Transformer, showing how the assignment goes back and forth between them.
A list of the languages that students presented on: Ancient Greek, Arabic, American Sign Language, Bahasa Indonesia, Bengali, Cantonese, Dutch, Esperanto, French, Georgian, German, Great Andamanese, Hindi, Inuktitut, Italian, Japanese, Korean, Maltese, Mandarin, Mongolian, Nahuatl, Nepali, Romanian, Sanskrit, Spanish, Swedish, Tamil, Telugu, Tshiluba, Unangam Tunuu, Welsh, Wolof, Yiddish
Top plot: o1 scores better on several tasks when the output log probability is high than when it is low. Bottom plot: o1 uses more tokens when the output is low probability than when it is high probability.
Plots showing OpenAI o1's performance on common and rare versions of tasks. In general, on our basic evaluations, o1 is close to 100% accuracy on both common and rare task variants. However, on harder evaluations, it shows some separation, with stronger performance on common task variants than rare ones.
A plot showing LLM performance on various algorithmic tasks. For all LLMs evaluated, including o1-preview, performance is highly influenced by the probability of the output to be produced, with lower performance on cases with low-probability outputs. The tasks being evaluated on are shift ciphers, Pig Latin, article swapping, and reversal.
An image of a table with four columns. The first three columns are passages from translations of the Odyssey (by T.E. Lawrence, Robert Fagles, and Emily Wilson). The last column is labeled "Honda" and shows a passage from the Honda Odyssey owner's manual.
Images of two paper titles: "Sparks of Artificial General Intelligence" and "Embers of Autoregression"
Top: syntax tree for the sentence "the doctor by the lawyer saw the artist" Bottom: a continuous vector
A credit card reader displaying the brand name "worldpay"